The legend goes like this. In the spring of 1562, the 16-year-old prince Don Carlos of Asturias, grandson of the Holy Roman Emperor and heir to the Castilian throne, lay dying. The prince had been chasing a maid down a flight of stairs when he fell and hit the back of his head. He was taken to bed, weak and feverish, and his condition quickly deteriorated. His wound became infected, his head became swollen and he grew delirious. News of his illness spread: prayers and fasts were organised and his father, Philip II of Spain, arrived at his bedside with physicians, nobles and prelates in tow. The doctors applied ointments and poultices and gave him strong purgatives. But after weeks of debate they decided to trepan his skull in order to relieve the pressure from what they thought must be a cranial fracture. The surgeons gouged the flesh around the wound but found the skull to be unfractured, and halted the operation in dismay. What could be done to help the stricken prince?
The same day, hours after the doctors had laid down their drills, a procession of Franciscan friars from the local monastery appeared outside Don Carlos’s palace. They carried with them the preserved remains of a former member of their order, Diego de Alcalá: a lay brother to whom numerous miracles had been attributed. On being granted entry by Philip, they laid Diego in bed alongside the prince and unwrapped the winding sheet from around the dead man’s head. Witnesses say that Don Carlos, then blind from the infection and passing in and out of consciousness, groped instinctively for the corpse, running his hands across its face before touching his own eyes and mouth and collapsing again. That night the prince slept soundly for the first time since his fall and the next morning his pulse was stronger. His delirium abated, the infection retreated and the doctors judged that the worst was over. On regaining consciousness, Don Carlos said that a vision of Diego had appeared by his bedside during his illness, assuring him of God’s protection.
In thanks for this intercession (and perhaps to encourage a little reflection on the part of the maid-chasing prince), Philip II had a marvel built: a clockwork automaton of Diego sixteen inches tall and crafted from wood and iron. When stood on a table and activated by a hidden switch under its robes, the mechanical monk walked in circuits, lifting a wooden cross in one hand while the other beat its chest in a mea culpa, the mouth opening and closing in silent prayer. A miracle book written to sponsor Diego’s canonisation claimed he overcame paralysis on his deathbed and that his corpse lay uncorrupted in a chapel for months, so healthy in appearance that one visitor took his pulse in disbelief. Now, the monk’s mechanical likeness, commissioned by a pious king for an impetuous son, would offer similar testimony to the power of saintly bodies. The automaton resides today in the Smithsonian, reminding us of humanity’s long-standing desire to create life that survives death – an impulse once fulfilled through religion, and now often through technology.
How true is the Don Carlos story? Certainly he fell, received a visit from the dead monk and recovered, but as Elizabeth King and David Todd detail, the supposed origin of the machine is supported more by circumstantial evidence than positive proof; it’s an ‘elegant hypothesis’, the authors conclude. More interesting than the clockwork Diego’s uncertain provenance, however, is the tantalising concept of pre-modern artificial life. When we think of early automata we usually think of the well-known examples created in the 17th and 18th centuries, such as the defecating duck and the life-sized musicians made by the French engineer Jacques de Vaucanson. These automata embody the era’s materialist and mechanistic philosophies, inspiring thinkers like Descartes and Leibniz and finding easy coherence with our own robot-building. But how were such machines perceived before the Enlightenment, in an era when the border between magic and mathematics was more ill-defined? Did the monk’s original audience believe that the machine was powered by technology, faith or something else? Was it a toy or a devotional object; was it met with laughter or piety? King and Todd point to the ‘paradox of a rational object performing a mystical act’, asking: ‘What can it reveal of the birth of the machine in a still-fervent age of faith?’
Todd, a clockmaker and conservator, carries out a scrupulous autopsy of the mechanical monk. Carved from poplar or beech, it has a hollow, bell-shaped body that contains its clockwork. Three wheels on the underside provide propulsion and steering while a pair of sandalled feet on the front paddle at the air to give the illusion of walking. The arms are articulated at the shoulder and elbow; the head has iron eyeballs that scan the room and the drop-jaw of a ventriloquist’s dummy. The body is mostly featureless, but the face is as vital as a portrait, with an aquiline nose, elegant brows and smile lines. And though the machine’s paint is now cracked and faded, Todd and King note that the monk was ‘clearly once aglow with life; a faint flush still lingers on cheek and lip.’ Derobed and immobile, it evokes feelings of lost grandeur. ‘Looking at you now is like/seeing a god or a king/naked and starving in a field,’ the poet Rebecca Morgan Frank wrote in ‘Monk Automaton, c.1560’ (2021), an ode to the machine. But clothed and activated the monk becomes uncanny, even whimsical. When placed on a table it traces the path of a seven-pointed shape. With its head turning left to right, and its eyes swivelling, it seems to seek out the gaze of audience members. Are you watching? Do you see? Just before it turns on the points of its star, the monk raises its cross to its mouth and, with a smack of its wooden lips, offers a kiss of devotion.
The heart of the choreography is a clockwork cam in the machine’s chest: a metal disc turned by a coiled mainspring that engages internal levers. These levers pull on chains that connect to the monk’s limbs, tugging them into action like bells in a belfry. One half of the cam disc’s circumference is occupied by six smoothly rising notches; a seventh is set apart and juts out like a tiny spur. These notches control the movement of the monk’s mouth, and if you imagine the cam’s perimeter unrolled like an audio waveform you can see the way the different shapes correspond to different motions. The six slow-risers lower the monk’s jaw in prayer while the spiked seventh creates the sharp smack of the kiss. The pins on the face of the cam that control the monk’s eyes are placed on the opposite side of the disc to the spur. This, as Todd and King note, is to ensure that ‘the eyes stay focused on the cross at that penultimate moment.’
Todd and King have gathered together a cohort of seven similar figures from museum collections in Europe and the US. Each is of comparable size to the monk, with a clockwork core, wooden body and mobile arms. These automata have been called ‘the cephalic androids’, thanks to the realism of their heads and the careful orchestration of their eyes. But Todd and King go one step further and argue that they ‘comprise a class by themselves in the history of artificial life: the first solo free-walking self-propelled androids’. The other figures offer variations on the monk’s theme. One is a lay brother worrying a tiny beaded rosary; another a bearded saint crowned by a halo, with a six-note glockenspiel in its rump including miniature bells and hammers. The rest are secular: young women holding lutes, citterns and mandolins. An account of one of these players by the art historian Fritz Saxl from 1936 describes the way it moves with ‘a combination of stateliness and grace … The feet move slowly, the figure seems to float. The skirt disguises the shape of the body completely. In contrast, the hands and head are in free and elegant motion. The head describes exquisite curves to the music produced by the sensitive delicate hands.’ Todd and King speculate that the choice of subjects – religious men and courtly women – may have been dictated by the shared technical blueprint of the automata: each machine needs some sort of curtain, robe or dress to hide its internal mechanism and create the illusion.
The mechanical monk does resemble Diego de Alcalá (or popular representations of Diego, who died in 1463). Its construction can be dated to the 16th century thanks to tooling marks on its clockwork, and details of the figure match the monk’s story, including the fully shaven head that is the mark of a lay brother rather than a tonsured initiate. We know that both Charles V and Philip II employed a clockmaker, the Italian-Spanish engineer Juanelo Turriano, who is said to have built other automata. Turriano was by all accounts a technical genius. The painter Antonio Campi described him in 1585 as ‘a low-born man, but gifted by God with such a sublime intelligence that he astounded the world’, while the poet Marco Vida compared him to the god Vulcan, ‘with his face, hair and beard covered and smeared with abundant ash and disgusting soot, with his thick and enormous hands and fingers always full of rust’. One of his great accomplishments was an astronomical clock, now lost, which took twenty years to design and which he inscribed: ‘QVI SIM SCIES SI PAR OPVS FACERE CONABERIS’ (King translates this as ‘you will know who I am if you try to make this’). A 17th-century history of the Low Country (part of Charles V’s domain) by the Jesuit historian Famianus Strada says Turriano frequently built automata for his royal patrons, to provide entertainment after meals:
When the Cloth was taken away after dinner, he brought upon the board little armed figures of Horse and Foot, some beating Drums, others sounding Trumpets, and divers of them charging one another with their Pikes. Sometimes he sent wooden sparrows out of his chamber into the Emperours Dining-room, that would flie round, and back again; the Superiour of the Monastery, who came in by accident, suspecting him for a Conjurer.
Such after-dinner spectacles were common in the 16th century for the rich and powerful; popular automata included wheeled ships that sailed through the clutter of serving dishes and cutlery, offering a spectacular reminder of wealth and reach. One ship, Hans Schlottheim’s Mechanical Galleon, has moving figures on the deck representing the Holy Roman Emperor Rudolf II and his prince electors, along with trumpeters, drummers and a set of miniature cannons that fire squibs as the ship moves – all powered by a hidden set of bellows in the galleon’s hull.
No automaton known to have been made by Turriano has survived, however; there is no maker’s mark or paper trail to show that he created the monk. And there are no contemporary accounts of the machine – which is odd, given the existing records of other figures and the monk’s memorable appearance. Todd and King find out little about its provenance until it enters the public record in the 1970s. In 1975 a Swiss antiques dealer, Georges Sedlmajer, approached a Franciscan friar and historian, Servus Gieben, to ask about the machine. It was Gieben who connected it to the legend of Don Carlos. It was acquired by the Smithsonian two years later, which is where King and Todd encountered it. Though they can’t verify its origin story, they show that the figure and the Don Carlos legend illustrate our broader fascination with trying to grant life to inanimate matter: ‘The monk models in votive intercession or sympathetic repair everything the physicians were trying to accomplish: to prevent the ebbing of life from the body, to rewind its vital spring, to put the prince back on his feet so that he might move again under his own power.’ Such a story captures the miracle implicit in the creation of all automata: ‘a motionless thing that comes to life’.
The historian of science Jessica Riskin has noted that the Church was the greatest sponsor of proto-robots in the late Middle Ages in Europe. Cathedrals and monasteries not only possessed the material wealth necessary to create mechanical spectacles but nurtured technical knowledge. It was the Church that translated and printed ancient sources of mechanical wisdom such as Vitruvius’s De Architectura and commissioned the most technologically advanced creations of the day: astronomical clocks. The religious interest in timekeeping stems from monastic schedules of prayer such as those described in the Rule of St Benedict. Keeping appointments was made easier in the late 13th century by the development of mechanical clocks, which stored energy in a hoisted weight or coiled spring and released it in a controlled, incremental manner. This mechanism was what served as the tick of a clock – or the movement of an automaton. Astronomical clocks became increasingly complex as designers strove to capture more of the divine schedule in gear and springs. Carillons, astrolabes and calendars were incorporated, marking feast days, phases of the moon and the movement of the planets. For these clocks to be useful, though, bells had to be rung to mark the hours and gather the penitent, so among the first proto-automata built in Europe were jacquemarts or quarter-jacks: human figures with mallets that struck the hours in churches and monasteries – a common sight across Europe from the mid-14th century.
The most celebrated example of such horological pageantry is the astronomical clock of Strasbourg Cathedral. Built between 1352 and 1354 and refurbished between 1540 and 1574 (the version visible today dates from the 19th century), the clock houses a huge cast of figures. The 16th-century version included not only a cockerel that flapped its wings and crowed on the hour, but representations of the three Magi, the Virgin and Child, a septet of Roman gods (one for each day of the week), an angel who turned an hourglass every quarter of an hour, a procession of twelve apostles who shuffled into view at noon to be blessed by a mechanical Christ and, centre stage, the skeletal figure of Death, acting as jacquemart with scythe in one hand and bone in the other to strike the hour. ‘Astronomical clocks showed the glory of the macrocosm; the automata on them embodied the microcosm,’ as Elly Truitt puts it in Medieval Robots (2015). Both served to animate and dramatise the story of ‘divine creation and human salvation’.
Mechanical angels and devils were standalone figures in some churches. At the simplest end of the spectrum were wooden statues with articulated but unpowered limbs, like the 13th-century Cristo de los Gascones, a life-size depiction of Christ with jointed arms that is preserved today in the Church of San Justo in Segovia. Such figures are puppets rather than automata, and were used in religious ceremonies that blurred the boundary between liturgy and theatre. In the case of El Cristo de los Gascones, the figure’s posable limbs allowed church members during Good Friday celebrations to re-enact the deposition of Christ. Church members would assume the roles of Nicodemus and Joseph of Arimathea, taking the wooden Christ down from a cross and moving him to a sepulchre on the other side of the church; a sort of theatre in the round.
Other devices were more complex, supplying their own motive power and so making them true automata. Angelic figures were often attached to organs (another locus of mechanical expertise in the Church), where they would sing hymns and raise trumpets, powered by the same pneumatics that fed the instrument’s pipes, while others could be placed in niches like statues. In The Restless Clock (2016), Riskin mentions a 16th-century devil carved from wood that appeared to burst from a cage: ‘horrible, twisted, horned, rolling furious eyes, sticking out a blood-red tongue, [it seemed] to throw itself upon the spectator, spitting in his face and letting out great howls.’ Inside was a set of hidden bellows powered by a suspended weight: the devil must have functioned like an American Halloween animatronic, frightening the unwary with a sudden burst of sound and movement. One of the most celebrated examples of ecclesiastical mechanicals is also one of the most complex: the Rood of Grace, an articulated sculpture of Christ that was housed in Boxley Abbey in Kent from the 15th century. The Rood would enact the saviour’s suffering for visiting pilgrims, writhing and grimacing on the crucifix above them. It’s not completely clear how the figure was powered, but according to an account by the 16th-century antiquarian William Lambarde it was able to ‘bow down and lifte up it selfe, to shake and stirre the handes and feete, to nod the head, to rolle the eies, to wag the chaps, to bende the browes’.
How did medieval and early modern audiences perceive such creations? Todd and King don’t spend much time on this question, perhaps because there are so few contemporary sources relating to the mechanical monk. Certainly, not all religious automata were treated with the same reverence as relics or devotional artwork. Some were intentionally funny, like the Bretzelmann, or ‘pretzel seller’, of Strasbourg Cathedral – a figure attached to the organ that waved its arm and bellowed ‘profane and bawdy songs in a raucous voice’ at the congregation below (with vocals provided by a hidden human performer). Similarly, the statue of a saint that could be raised and lowered with trick machinery was said to have produced ‘more laughter than devotion’. Did people think the automata were ‘real’? How did they imagine they worked? Such questions are natural but too grounded in a modern conception of technology. Spectators would perhaps have known that the figures were mechanical, but this was a category with very different boundaries from those of today.
Our first misunderstanding would be to place stories of ancient automata in an unbroken lineage that leads to the present day, as if humans’ conception of artificial life were continuous. An example of an ancient ‘automaton’ often cited is Talos, the thirty-metre brass giant of Greek mythology, crafted by Hephaestus to protect Crete. Talos is quite unlike a robot, more living metal than machine man. According to Apollonius Rhodius’ Argonautica, an epic from the third century BCE, ‘his body and his limbs were brazen and invulnerable, except at one point: under a sinew by his ankle there was a blood-red vein protected only by a thin skin which to him meant life and death.’
Medieval fiction often conceptualised artificial beings in a similar way: constructed from inanimate matter and mimicking aspects of human behaviour, but not solely mechanical creations. Take the gigantic copper knights that patrol the formidable castle known as the Dolorous Garde in the Lancelot du Lac prose cycle, which are disarmed when Lancelot finds their power source: a box of evil spirits trapped in copper tubes. Or the oracular brazen heads supposedly created by Albertus Magnus, Robert Grosseteste and Roger Bacon. These heads were forged from bronze or brass and activated with the help of arcane knowledge of spirits or celestial conjunctions. Once operational, they were able to predict the future and answer any question. But stories about the heads’ creation usually end in them being destroyed, by accident or divine intervention – or speaking only when their master has left the room or fallen asleep from exhaustion. These tales were widespread and functioned as fables on the dangers of forbidden scholarship that might trouble the authority of the Church.
As the brazen heads demonstrate, artificial life could be both mechanical and magical. The term ‘automaton’, referring to a self-moving machine, was first used in Europe in 1531 in a catalogue of magic, De Occulta Philosophia, compiled by the Hermetic philosopher Heinrich Cornelius Agrippa von Nettesheim. Automata are listed as a type of ‘celestial magic’, a category which includes maths, music, astronomy and mechanics. Study these arts in combination, Agrippa writes, and your knowledge will allow you to create spectacles of light, motion and harmony, including artificial animals and servants. Another point of triangulation is the concept of the preternatural, promulgated by Thomas Aquinas and described in an essay by Lorraine Daston as ‘that twilight zone between the natural and the supernatural’. As Daston notes, the preternatural was a particularly influential idea in the High Middle Ages – uncanny and marvellous, positioned between things as they normally are and events attributable to God.
In the case of religious mechanicals the picture is further complicated by theology; in particular, a pre-Reformation understanding of image and representation. As Riskin writes, a core tenet for religious reformers in the 16th century was the refutation of the Eucharist and the miracle of transubstantiation. Until this intervention, Riskin argues, religious automata would have existed in a different conceptual universe, one in which icons and images didn’t just portray Christian belief but manifested it. ‘Rolling their eyes, moving their lips, gesturing and grimacing, these automata dramatised the intimate, corporeal relation between representation and divinity, icon and saint … The icons were mechanical but neither passive nor rote. On the contrary, they were representations in motion, inspirited statues: they were mechanical and divine.’
The inclusion of automata in this theology can be seen in the post-Reformation treatment of the Rood of Grace. Henry VIII banned automata from England’s churches as part of the creation of the Anglican Church, and the Rood was taken from Boxley Abbey in 1538 at the time of the monastery’s dissolution. The removal was carried out by Geoffrey Chamber, an agent of Thomas Cromwell, who reported to his master that he had plucked down the Rood, ‘which has been had in great veneration’, only to find ‘certain engines and old wire, with old rotten sticks in the back of the same which caused the eyes to move and stir in the head thereof, like unto a lively thing’. Chamber wrote that he confronted the abbot and monks but the men ‘declared themselves ignorant of it’. In order to disenchant the locals, he sent the Rood to nearby Maidstone – where, in the words of the chronicler Charles Wriothesley, members of the public were able to see for themselves ‘the craft of movinge the eyes and lipps’ by which the monks ‘had gotten great riches in deceiving the people’. Finally, the Rood was transported to London, where it was displayed before a crowd at St Paul’s Cathedral like a captured king. There it was the subject of a sermon by John Hilsey, bishop of Rochester, who denounced the ‘great adolatrie’ of the monks of Boxley and the ‘abuses … of engines’ before encouraging his audience to break the Rood into pieces and burn it.
How much of this context can be applied to the mechanical monk is uncertain, but it’s clear that automata often had a hieratic function – inducting audiences into belief in higher powers, religious or otherwise. Humanoid robots today can occupy a similar role, enticing audiences with the possibility not of divine miracles but of otherworldly and fabulous technological futures featuring super-intelligent computer brains and legions of tireless artificial workers. Consider two of the best-known humanoid robots built in the past decade: Sophia, built by the Hong Kong-based Hanson Robotics, and the Tesla Bot, known as Optimus. Sophia has appeared on TV shows, magazine covers and the international conference circuit, answering questions about the future of technology and lending a sprinkle of glamour to otherwise staid developments. As with the monk, great care has been taken with Sophia’s appearance. It’s a lifelike mannequin with an exposed robot ‘brain’, similar to the lead robot in the 2014 sci-fi movie Ex Machina. (Sophia’s creator, David Hanson, once designed animatronics for Disney’s theme parks.) On stage, the robot’s handlers often exaggerate its capabilities, with Hanson in one interview even describing it as ‘basically alive’.
Others involved in Sophia’s creation are more candid about its purpose and utility. Ben Goertzel, Hanson Robotics’ former chief scientist, has described it as a hierophant, introducing people to belief in artificial general intelligence or AGI. ‘If I tell people I’m using probabilistic logic to do reasoning on how best to prune the backward chaining inference trees that arise in our logic engine, they have no idea what I’m talking about,’ Goertzel told me in 2017. ‘But if I show them a beautiful smiling robot face, then they get the feeling that AGI may indeed be nearby and viable.’ Goertzel notes that Sophia’s non-stop publicity tour acts as a counter to the marketing heft of better-funded competitors, who prefer to use more subtle pageantry. ‘What does a startup get out of having massive international publicity?’ said Goertzel. ‘This is obvious.’
In 2021, Elon Musk unveiled his own humanoid robot – or rather, he unveiled a man in a spandex robot costume who danced on stage to dubstep before the billionaire presented a pitch deck for a working machine. Since this ‘unveiling’, Tesla’s engineers have produced several working prototypes of Optimus. But while experts say the company has worked quickly, it’s unlikely that the machine will match the utility of human workers any time soon. Androids are notoriously difficult to build, with engineers forced to compete with millions of years of human evolution, in terms not only of biological hardware (muscle density, power storage) but also of software (balance, eye-hand co-ordination). They do offer certain advantages, such as the ability to slot seamlessly into working environments built for humans. But robot-shaped robots are more practical. The most effective robot in your house is a waterproof box filled with jets and sprays rather than a pair of mechanical arms that attempts to wash the dishes for you. Some humanoid robots can put on a good show: moving heavy objects, climbing ladders and doing gymnastic routines. But they are still experimental, and it’s unclear how many decades it will take for the robot worker that Musk promises – able to perform any task a human can – to become a viable prospect.
Why is Tesla developing Optimus when its funds could deliver more immediate benefits elsewhere? The bot’s presentation encapsulates Musk’s methods in selling his vision: bombastic and confrontational, a blend of the visionary and the delusional. As ever, it’s all about the spectacle – about harnessing powers that exist somewhere between the mechanical and the magical. When videos of Sophia or Optimus are shared online, people respond with fear, delight and awe. Are they being fooled? Do they believe?
Send Letters To:
The Editor
London Review of Books,
28 Little Russell Street
London, WC1A 2HN
letters@lrb.co.uk
Please include name, address, and a telephone number.